应用双树复小波包和NCA-LSSVM检测磁瓦内部缺陷

发布时间：2018-05-27 19:03

本文选题：磁瓦 + 内部缺陷　；参考：《浙江大学学报(工学版)》2017年01期

【摘要】：提出结合双树复小波包变换(DTCWPT)、邻域成分分析法(NCA)、最小二乘支持向量机(LSSVM)的磁瓦内部缺陷检测方法.通过双树复小波包将采集的声音信号分解为6层,得到64个不同频带的子信号;求取特定频带信号的能量、偏度、峭度、模糊熵,并将能量、偏度、峭度、模糊熵作为分类特征;利用邻域成分分析法对分类特征降维;将降维构造的新特征集输入到最小二乘支持向量机,判断磁瓦是否含有内部缺陷.通过实验验证,对提出的检测方法进行可行性分析.3种不同类型磁瓦的内部缺陷识别率均可以达到99%,与以往双谱切片方法相比,提高了检测识别率.试验结果表明,提出的方法具有检测速度快、可靠性高、适应性强等特点,为高效、准确地进行磁瓦内部缺陷检测提供了有效的技术手段.
[Abstract]:In this paper, a new method for detecting internal defects of magnetic tile is proposed, which is based on double tree complex wavelet packet transform (DTCWPTT), neighborhood component analysis (NCAA) and least square support vector machine (LSSVM). The collected sound signal is decomposed into 6 layers by double tree complex wavelet packet, and 64 sub-signals in different frequency bands are obtained, and the energy, skewness, kurtosis, fuzzy entropy, energy, deviation, kurtosis of the signal are obtained, and the energy, skewness and kurtosis of the signal are obtained. Fuzzy entropy is used as the classification feature; neighborhood component analysis is used to reduce the dimension of the classification feature; and the new feature set of dimensionality reduction is input into the least squares support vector machine to determine whether the magnetic tile contains internal defects. Through the experimental verification, the feasibility of the proposed detection method is analyzed. The internal defect recognition rate of 3 different types of magnetic tile can reach 99%. Compared with the previous bispectral slice method, the detection and recognition rate is improved. The experimental results show that the proposed method has the characteristics of high detection speed, high reliability and strong adaptability, which provides an effective technical means for the efficient and accurate detection of the inner defects of the magnetic tile.
【作者单位】：四川大学制造科学与工程学院;
【基金】：国家自然科学青年基金资助项目(51205265) 四川省科技计划资助项目(2015ZR0018)
【分类号】：TM273

【相似文献】